Opencl cpu batch matmul fix #2774
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
d94382d to
a4eb27a
Compare
The Intel OpenCL mapping the same buffer for write access caused an error. This caused the matmul operation to fail when the same array was passed in. To fix this only the READ flag is passed into the map function instead of the READ and WRITE flags
The new getMappedPtr is now (correctly) returning a unique_ptr. This commit removes implicit conversions from unique_ptr to shared ptr.
a4eb27a to
2356ec9
Compare
9prady9
reviewed
Feb 21, 2020
9prady9
reviewed
Feb 21, 2020
9prady9
approved these changes
Feb 21, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes a couple of issues with the CPU OpenCL blas functions. The batching functionality was not working correctly for righ hand side batches. The matmul function was throwing errors in case the same input was passed in. This was caused only on the Intel GPU OpenCL implementation because mapping the same buffer as a READ and WRITE mapped pointer caused an error. These issues have been fixed and the getMappedPtr is refactored to return unique_ptr instead of shared_ptr.
Fixes #1711